Graham McNicoll

Most engineering teams using feature flags treat them as circuit breakers. Something goes wrong in production, flip the flag, turn it off. That is genuinely valuable, and most engineering teams think about flags exactly this way. What I find interesting is how close they are to getting much more out of the same system. The hard work in feature flagging is the integration itself. Wiring flags into your code, shipping behind them, building the muscle across the team. Those teams have already done the heavy lift. The experimentation layer comes for free on top of that. The metrics are already tracked. Users are already being assigned. The hard part is done. What is left is attaching an experiment to the flag you already have. The same circuit breaker now tells you whether the feature moved the metric. If you already have flags in production, turn the next one into an experiment. And if your platform doesn't make this stupidly easy, give growthbook.io a go.